So with that drill if the client wanted to change the battery or another component - I find it a lot easier to accommodate changes in parametric software. Again - if we decided to change the material the housing will be molded with that required a different wall thickness - this is rather trivial in parametric software. I’d like to see this dealt with in Rhino where you need to update everything manually.
Our time spent designing is relatively quick so for most models there’s an initial time investment building a robust parametric model so that changes & iterations further down the line are quite quick to implement. If I had to rebuild surfaces and manually move features in Rhino every time a change was made I’d spend far too much time doing (unnecessary) CAD work.
You certainly can do the housing in rhino and components in fusion, but for me it comes down to workflow and how quickly I can make design changes and for that reason I prefer to have everything done in my primary cad software.
But by all means explore using Rhino + fusion - it might work really well for your particular workflow.
Ive been trying to do thje same thing the author is asking for for 2 yrs now. There isnt a lot of videos or tutorials to go from start to finish. However, Ive found there are videos, that go point A to B, B to C, and sometimes have to figure out on your own from C to D to get the finished product. Sometimes frustrating , sometimes satisfying.
This is a double edged sword, because if everyone else figured it out, there would be not as much need for what the author and myself is trying to do.
youre right. this seems to be a pretty guarded process. I’m determined to figure it out though and i think im in the right place. At some point i think i am going to have to pay for some sort of training tutoring seminars or classes to get what i really want. but up until im sure what that is im going to keep trying to figure it out. If you know any useful tutorials please share
I don’t think there’s anything guarded about it, probably more just a case of it being a laborious process of creating new geometry that closely matches your scanned geometry.
Process used by an individual can depend on:
Geommetry
Details of the scan including accuracy, completeness, number/density of points
Use of the result
Experience and knowledge of the individual
There is no single process for every situation. What works for one may completely fail for another. Or may be much more laborious than needed.
I don’t think anyone is trying to keep the reverse engineering process a secret. If one is using a tool that is not designed for reverse engineering (out of the box Rhino), the process is going to be a lot more problematic that if one is using DesignX, Mesh2Surface, MeshToCAD or a similar tool. These are not easy applications to learn and use, but they provide tools that are designed with the reverse engineering process in mind.
If the large part one one’s time is spent doing reverse engineering it would make sense to pay for a tool and the training that would save significant amounts of time, if the budget supports it.
Mesh2Surface (Rhino Plugin) and Cyborg’s Mesh To CAD (standalone) are “affordable” options compared to DesignX. Users here who are feeling the pain of reverse engineering projects may want to investigate.
@cdordoni Do these tools produce a smooth NURBS surface when used for objects such as auto interiors and exteriors, boat hulls, many consumer products, etc. Or are they primarially for use with shapes comprised of planar, cylindrical, conical etc surfaces?
You can get smooth organic surfaces as nurbs from the scan with these tools. And they can work with many more triangles in the scan than Rhino’s own tools.
EDIT - I just wanted to add that there may be more nurbs patches that are produced than the ideal, very similar to Rhino’s subd tools, if you are using the more or less automated patch generation.
If you are looking for something with fewer patches then you might have to edit curves based on sections through the mesh. You would really have to test each software to see what’s possible with regard to editing curves.
I recommend 1 of the tools cdordoni mentioned in between the scan and Rhino.
Yes, I got lucky and had a basic version of Design X packaged with my scanner. That has been VERY helpful in reverse engineering. Repair scan mesh, fill holes, decimate, defeature, smooth scans, and even has its own version of a wrap just to name a few. I do a lot of cross sections and export as curves, export mesh, and import both into Rhino for rebuilding. I can use mesh for reference, and manipulate curves to match mesh.
That’s “one way” of doing it. There’s other ways that others have described above. Without RE specific software, the processes that work would be slightly more meticulous upto extremely more meticulous, depending on design intent.
I shared a project I did where I used shrinkwrap, subd, and nurbs to engineer a new part from a 3D scanned part. My design intent was only like +/- 0.15625" from original – if not more.
So the more accurate you have to RE, the more difficult and meticulous all the operations entail – of course.
What is it exactly?
Which is?
I’m not understanding what idea to which you’re referring exactly. Some type of particular ‘workflow’?
Almost forgot, yes Rhino is pure freedom. However there are always pros and cons in any scenario, so with pure freedom we don’t get to have ‘constraints’ yet, cause the developers are too cool
They like going into the abyss with GH and all that parametrics, but they totally missed the basics of parametric solid modeling that all the other CAD’s do. Simple stuff that’s ‘blatantly’ been ‘avoided’ – so far. Although there’s signs of hope.
I’ve heard this saying before, but I’m not understanding it atm, although I could google it or somem
They say GH is parametric.
I’d like to see everything done in Rhino, and everything as automatic as possible.
That’s one of the big ideas behind parametrics as I’ve always understood it. Things down the line can indeed theoretically become very complex and very helpful – depending on the design intent of things that should be compatible with such parametrics.
While, even very simple stuff immediately ‘in the line’ would be helpful, such as “sketches” that can be geometrically and dimensionally constrained in real time as they’re being created.
Having that in Rhino would be amazing, but Rhino should use the ‘3D all the time’ approach to this, because that’s the nature of Rhino. And Rhino should approach this avenue naturally in it’s own unique way, rather than the boring common way of the other CAD’s.
I’ve used so many other CAD’s that have very gross ways of referencing 2D, 3D sketches – if at all. Usually they were incapable of it – on the low end. I’m sure in modern times they do this easily … right…
ikr! , but you do get really good at using Rhino tools in the process clicking so many buttons repeatedly
That’s why we need ‘constraints’ options in Rhino
Yeah my opinion is why use fusion at all? It makes no sense to me. Their license terms are terrible, and their cloud based tech is a breach of security. Just makes zero sense. My algorithm goes fusion?-‘nope’.
Yeah, and magicians don’t usually reveal all their tricks.
Indeed.
I’m still tryna figure out what workflow you’re after. If it has something to do with fusion then that’s the problem imo. Stop using fusion lol.
If your reason is cause you don’t like modeling stuff over and over and over in Rhino, well technically that’s just lazy – to be blunt. After modeling 3D for 20 yrs, that mental block goes away. I have to model it again? Yes, model it again lol.
But in the meatime make threads in the forum and tell the devs to build us ‘constraint’ based options.
Well these days, yes you can find all the secrets much quicker than 20 yrs ago. But, there’s still not many users giving away all their tricks. Plus there’s no many options for affordable software solutions either.
Well, maybe true, but it’s like one of those hidden in plain sight type of things. You have to know what you’re looking at to see it.
Well I’m not a fan of the whole ‘training fee’ of most technology. Big companies love charging customers for “training”.
Ok so maybe saying it was gaurded was a hasty choice of words. Use case is surely a factor. So for whatever it’s worth. I’m designing interior trim panels and the like and mostly I want to use the models to Boolean bodies. Door panels and trims. Center consoles and various features inside of a car. That kind of stuff. I figured a lot of this could be done with surface modeling like in rhino. I’m using a einstar scanner right now and I’m getting better. And my plan is to get a better scanner when it makes sense. There’s a lot of good information here but am I on the right track?
I have “better scanners”, and it’s kinda a long story, but I believe that scanner you got is adequate for most things, although I still need to get one and mess with it.
Yes I’m pretty sure.
Yes. Just get your geometry from the scanner software into Rhino.
Correct me if Im wrong and my understanding of reverse engineering in 3d…
I come from a graphics background (2D like Illustrator, Photoshop, Corel, etc) where I realized I couldnt just convert a raster to a vector the same way as a mesh to NURBS… at least without being less than accurate. You will have to almost “trace” much of the final product.
I want a 3d scanner that scans perfectly into NURBS…lol
Technically, scanner tech today is pretty close, if you know the secret workflows. But the scan quality is directly proportional to the skill of the person handling the scanner and the data processing.
But everything I’ve witnessed over the last 20 yrs, I believe it’s possible. It will just take the right collection of individuals to put the technology together.
Rhino’s shrinkwrap command is hands down one of the best tools I’ve seen in the last 20 yrs in all the CAD’s and RE’s I’ve used. I’m sure I’ve used very similar tech, but Rhino’s I like the most because it has some Rhino flavor to it, whereby it has that freeform feel.
I’ve seen a multiple other softwares that do similar wrapping but they didn’t really let the user have an intuitive connection to the wrapping exactly. Idk, maybe I just like doing it in Rhino alot lol.
In some ways the subd’s kinda make nurbs obsolete, but not really. It really adds a complement format between mesh and nurbs. It’s a true trio link flow now. I enjoy using subd’s to go back and forth between meshes and nurbs in certain scenarios. It’s not automatic for every situation yet but it’s pretty fun using the flow. I guess it works best for organic geometry.
My favorite part about subd’s is how easy it is to run the smoothing command on selected points and get the geometry to smooth out. Meshes and nurbs just don’t have that same ease of use. While you can convert back and forth from subd to mesh, and then subd to nurbs if you need the nurbs.
So, reverse engineering is simply the measurement of points and the understanding of how things work. With that simple nature, that is RE. Now, if you use lasers to measure billions of points, then you don’t necessarily have to understand how something works if you measure billions of it’s atoms locations.
You might need to measure more than a billion atoms though, cause them are super small haha.
But yeah RE is all about capturing geometry, so “trace” is kinda a way to look at it per say.
With the possibilities in 3D printing, NURBS become less important…
And scan post processing software like Artec studio lets you fit basic NURBS like cylinders, spheres, tori and planes to scans. In addition to that these CAD geometries can be constrained so for example two holes can be defined parallel or perpendicular to a face just to make an example…
All scan results as NURBS I don’t think makes sense. You’ll have to learn how to work with meshes.
This is like +/-0.005" to the mesh. To improve precision, the mesh quality would need to be better from the start.
If I were to guess, the mesh I’m seeing is probably like +/-0.15625" from the actual part from the start. But I’m sure there’s better meshes somewhere, otherwise it would be wise to adjust the scanner workflow parameters prior to importing into Rhino.
Design intent is key though. It’s possible the deviation allowance is 0.25" or more.
you’re right for saying the tolerance could be .25. its actually whatever i want it to be. i would ideally like to be in the .005 range though. and yes the scan could be better. i haven’t exactly mastered scanning yet. But i opened the file you posted and wasn’t exactly sure of what i was looking at. I’m still a bit new to rhino here. you said you used the “pull” command? are you “pulling” a NURBS surface “down”? Ive heard of this as “pulling down.” But you say you would rather use nurbs or sub D. im just a little confused and couldnt really tell what you had done in the file. I had used subd and with a little time i could have gotten it closer if i had inserted more lines and “pulled” vertices into position a bit better. All i really needed was a smooth surface that would translate well to a 3d printer and fit close to perfect in real life. what are you doing that i wasnt?
, but you do get really good at using Rhino tools in the process clicking so many buttons repeatedly 
